Beyond “Environment → Outcome”: What taking the “Person × Environment” perspective seriously changes in research practice

external determinants of measurable psychological responses. However, the growing emphasis on measurement, psychometric refinement, and explanatory models centered on environmental predictors has progressively shifted attention away from the discipline's original theoretical premise: that psychological experience emerges through ongoing transactions between people and environments. In this Opinion, we argue that this evolution has encouraged an implicit Environment → Personal (Psychological) Outcome perspective in which places are primarily conceptualized as external determinants of measurable psychological responses. Although this approach has generated substantial empirical advances, it risks reducing inherently relational phenomena to linear associations betweenEnvironmental psychology was founded on a distinctive theoretical premise: psychological experience, behavior, and well-being emerge from continuous transactions between people and their physical and socio-cultural surroundings (Proshansky et al., 1970;Stokols, 1978).The premise has deep roots; in Lewin's (1936) formulation of behavior as a joint function of person and environment, in Barker's (1968) ecological psychology, and in the transactional worldview articulated by Altman and Rogoff (1987), which treats person and setting as mutually constitutive rather than analytically separable. Unlike frameworks that conceptualize environments as external conditions producing measurable effects on individuals, the discipline has traditionally emphasized the reciprocal, contextual, and temporally extended character of human-environment relationships (Bonnes and Secchiaroli, 1995).We do not offer P×E as a new framework. Previous theoretical perspectives already treat person-environment relations as reciprocal, contextual, and dynamic (Altman & Rogoff, 1987;Barker, 1968), and Ruotolo and Aletta (2026) likewise argue that context is constitutive of, rather than a background moderator to, psychological processes. Our contribution is of a different kind. Ruotolo and Aletta's aim is theoretical integration across mechanisms and scales; ours is narrower and complementary: the study-level logic by which the premise is translated into decisions about design, sampling, modeling, and interpretation, including the sampling of environments and the treatment of heterogeneity.These implications are domain-general and can be applied within, rather than instead of, existing domain-specific frameworks.Crucially, from the transactional this perspective, the "person" is conceptualized as the outcome of a multiplicative, rather than additive, interplay between nature and nurture.Both are constitutive of the person: nature refers to the genetic and biological endowment, while nurture encompasses all non-genetic and non-biological influences, including physical, ecological, social, cultural, and economic environments. These environments are therefore not external factors acting upon a separate individual; through development and ongoing interaction, they become embedded in and constitutive of the person. In this sense, the physical-ecological environment, including both built and natural settings, is a specific dimension of nurture, and the one specifically foregrounded by contemporary Environmental Psychology.For the purposes of this Opinion, we describe as Environment → Personal (Psychological) Outcome (E→PO) a recurring analytical tendency: treating environmental characteristics primarily as independent variables expected to produce measurable psychological consequences. We observe it most clearly in research on restorative environments, green space, and environmental health (Section 2). It is not characteristic of the field as a whole (Ratcliffe et al., 2026): other strands, such as work on place attachment and place identity, are explicitly relational.For the purposes of this Opinion, we describe this tendency as an Environment → Personal (Psychological) Outcome (E→PO) perspective: an analytical approach that treats environmental characteristics primarily as independent variables expected to produce measurable psychological consequences. The label is deliberately schematic, referring to a widespread analytical tendency rather than to any specific research program. This perspective has substantially advanced scientific understanding, but it may also distance the field from its foundational Person × Environment (P×E) perspective, in which psychological processes emerge from the dynamic interplay of people, places, and their histories, identities, relationships, and circumstances. This perspective has substantially advanced scientific understanding, but where it predominates it may distance research from the foundational Person × Environment (P×E) perspective, in which psychological processes emerge from the dynamic interplay of people, places, and their histories, identities, relationships, and circumstances. The issue, therefore, is not environmental measurement or quantification per se, but the reduction of complex environments and person-environment relations to isolated predictors and outcomes.We argue that important strands of the field face a tension We argue that the field now faces a tension between methodological refinement and the relational complexity that originally defined Environmental Psychology. Rather than choosing between E→PO and P×E, the next phase should integrate quantitative measurement within multilevel and longitudinal designs that situate psychological outcomes within the broader ecological and relational contexts.Accordingly, a specific aim of this Opinion is to draw out concrete methodological implications for empirical research in environmental psychology, translating the P×E perspective into actionable guidance for research design. Here, the '×' denotes a transactional, mutually constitutive relationship between person and environment, rather than a formal statistical interaction term, although the two senses converge in the moderation models discussed in Section 4.One of the most productive contributions of the field has been to demonstrate that physical environments shape human functioning. Research on natural settings, urban green space, restorative environments, and environmental quality has shown that specific environmental characteristics are reliably associated with psychological benefits, including reduced physiological stress, improved directed attention performance, and enhanced well-being (e.g., Ulrich et al., 1991;Hartig et al., 2014;Bratman et al., 2019;Spano et al., 2023;Ricciardi et al., 2026). Part of this evidence now supports approximate dose-response formulations linking weekly time spent in nature to self-reported health and well-being (White et al., 2019).Environments do not, however, act on individuals uniformly or deterministically. A place is not merely a stimulus that elicits a response; it acquires psychological significance through the encounter between its physical properties and the person who engages with it. Settings offer affordances that are realized only in relation to the capacities, goals, and states of a particular perceiver. Accordingly, between-person heterogeneity in environmental response is not measurement noise but a substantive feature of the phenomenon.A forest, a park, or a familiar landscape is not inherently restorative. It may be experienced as restorative because it affords reflection, evokes positive autobiographical memory, signifies safety and belonging, or meets a psychological need at a particular moment (Korpela, 1996;Korpela et al., 2001). Conversely, the same environment may be experienced very differently by individuals differing in history, identity, cultural background, or present circumstance. Place meanings are neither given by the physical setting alone nor freely constructed independently of it (Stedman, 2003).The distinction matters especially in contexts defined by climate change, environmental uncertainty, and the rapid transformation of familiar places. In this case, environments cannot be understood solely through their measurable effects on well-being; they are also sites of identity, attachment, memory, conflict, and adaptation (Adger et al., 2013;Tschakert et al., 2017). In this sense, the constructs of place attachment, i.e., the affective bond individuals develop with specific settings (Scannell and Gifford, 2017), and place identity, i.e., the incorporation of place-related meanings into the self-concept (Twigger-Ross and Uzzell, 1996), have become central to understanding how people relate to threatened or changing environments.The task for environmental psychology is therefore not only to identify which environmental characteristics produce benefits, but to explain how, for whom, and under what conditions environments become meaningful psychological resources.The development of reliable and valid instruments is among the discipline's major achievements. Psychometric work has allowed researchers to operationalize complex concepts and to investigate environmental experience with growing rigor. Yet increasingly refined measures also create a conceptual risk: research questions may come to be shaped by what is readily quantified rather than by the structure of the phenomenon under investigation.Perceived restorativeness is an instructive case. Attention Restoration Theory and subsequent work on restorative settings have substantially advanced understanding of environmentally supported cognitive recovery (Kaplan and Kaplan, 1989;Kaplan, 1995), and the Perceived Restorativeness Scale (Hartig et al., 1997) has been used extensively to assess the restorative qualities attributed to different environments. The issue is not the validity of such instruments: they are necessary building blocks of rigorous environmental research. Rather, they must be situated within a broader transactional framework.The broader question concerns which part of the person-environment relation is captured when restoration is measured in isolation from the wider fabric of experience. A restorativeness score describes how a respondent evaluates environmental qualities associated with recovery; it does not, by itself, indicate why a place matters to a particular person, how prior experience conditions present perception, or how identity, autobiographical memory, and social relationships participate in the experience of that place. Critical appraisals of the restorative-environments literature have raised parallel concerns about the specification of the underlying mechanism and about the considerable heterogeneity of effects across studies (Ohly et al., 2016;Joye and Dewitte, 2018). This becomes especially consequential in periods of environmental crisis. When people experience ecological loss, climate-related threats, or the disruption of familiar landscapes, psychological responses cannot be reduced to restoration or exposure effects alone. Places may simultaneously signify comfort and vulnerability, attachment and loss, identity and uncertainty (Albrecht et al., 2007;Cunsolo and Ellis, 2018;Clayton, 2020). The challenge, accordingly, is not to displace existing measures but to situate them within frameworks that preserve the relational character of environmental experience.Moving beyond the accumulation of isolated variables toward an integrated account of person-environment processes is, in our view, a central task for these strands of researchthe central task ahead. Many contemporary environmental problems show that human relations with environments involve multiple interdependent dimensions. Climate adaptation, environmental migration, biodiversity loss, and urban transformation are not merely physical processes; they are psychological and social experiences involving attachment, meaning, belonging, and identity (Adger et al., 2013;Tschakert et al., 2017).A strictly Environment → Personal (Psychological) Outcome approach can establish whether a given environmental feature is associated with a psychological outcome. It is less well equipped to explain why people form intense attachments to threatened places, why communities resist relocation despite documented risk, or why environmental change may generate distress and resilience simultaneously.Addressing such questions requires acknowledging that environments are not only contexts within which psychological processes occur, but they are constituents of those processes. This does not imply rejecting quantitative or experimental research but calls for methodological alignment with the relational complexity of the questions being asked.A few methodological implications follow.First, understanding how environments affect us requires the explicit modeling of mediating variables. For example, the effect of an objective natural environment on wellbeing may operate through cognitive appraisal and emotional self-regulation processes activated by the individual in a specific context.Second, understanding for whom and under what conditions environments matter demands the inclusion of moderating variables. Person-level variables, such as biographical history, cultural identity, current needs, pre-existing place attachment, situational circumstances, or perceived climate-related threat, should be specified a priori as constituents of the relational hypothesis, rather than introduced post hoc as background covariates. should be specified a priori as part of the hypothesis rather than introduced post hoc as statistical covariates.Taken together, these implications translate into four concrete research decisions.In design, person-level moderators and time are specified a priori and observed longitudinally or in situ, rather than averaged out. In sampling, both people and environments are sampled, with multiple exemplars per environmental category. In modeling, environmental features are not treated as predictors with person variables as covariates; people and environments enter as crossed random effects, with random slopes and moderators or mediators specified as hypotheses. In interpretation, effects are conditional on the sampled persons and places, and variance components are reported and treated as findings rather than as residual error.Third, environmental measures should be interpreted as indices of a relation rather than as fixed properties of settings, with corresponding caution in generalizing findings across populations and places.A further implication concerns the sampling of environments themselves.This concern extends to the sampling of environments themselves. Just as adequate sample size and power analysis are used to ensure that findings generalize across individuals, comparable attention is rarely paid to the environments under investigation: hypotheses are often tested using a single exemplar of a given environmental category (e.g., one park, one urban street), and effects are then generalized as though the category itself, rather than the specific setting sampled, had been tested.Treating environments as fixed rather than as one of several possible instances risks conflating the properties of a particular place with the psychological processes attributed to category membership. A relational P×E approach therefore also calls for sampling multiple exemplars within each environmental category, ideally within designs that treat environments and persons as structurally comparable, crossed sources of variability.Research designs should therefore capture time and context dynamically through longitudinal panels, in-situ experience sampling, participatory inquiry, and multilevel models capable of representing persons structurally nested within places.Treating environments as a sampled source of variability has a direct statistical consequence. With one exemplar per category, the effect of "green space" cannot be separated from the distinctive features of that place, and uncertainty about generalization to the category is understated (Clark, 1973;Judd et al., 2012). Only when multiple people encounter multiple environments can person, environment, and person-by-environment variance be estimated separately (Baayen et al., 2008). This reframes heterogeneity: variance in the environment effect across persons or across places is not residual noise but an empirical signature of the relational premise, and explaining it, through the moderators discussed above, becomes a primary aim rather than a robustness check (Bryan et al., 2021).For instance, a study on the restorative effects of urban green space could sample multiple parks varying in size, vegetation density, and perceived safety, treating these features as continuous, structurally comparable indices rather than as a single undifferentiated "green space" category, and modeling park identity as a crossed random factor alongside participants.In empirical research, the search for robust associations between environmental features and psychological outcomes often risks falling into the trap of "false correlations".The role of socio-economic status (SES) illustrates how the relational premise changes the modeling and interpretation of a single variable.Research frequently reports associations between greener neighborhoods and better mental health (e.g., Van den Berg et al., 2015, for an influential review). Income can play different roles in such associations, each corresponding to a different causal model. Under a selection model, higher income increases both the likelihood of living in a green, highquality area and mental health through other pathways; income is then a confounder of the environment-outcome association, and covariate adjustment is appropriate for estimating the average association net of selection. Under a relational model, income determines which environmental affordances are available and what a given setting means to the person: the same green space may be a basic survival resource for a low-income resident and a lifestyle amenity for someone with greater financial availability. Income is then an effect modifier, and the environment-outcome association differs across income levels.The two roles are not interchangeable, and they are not mutually exclusive: the same variable can confound the main effect and modify its magnitude, and a model should state which role each term plays. The standard response, including income as a covariate, answers only the first question. It returns a single average effect and assumes that effect is the same at every income level, so income-related heterogeneity is treated as noise to be subtracted rather than as part of the relation under is also only income is not a of the environments income (e.g., through income be a and for it part of the effect of a form of role is a matter of causal to be rather than of statistical relational advanced concerns the Income should not be treated only as a variable to be but also as a of the person-environment which we as a moderator of environmental We do not it as a of the green Accordingly, models should adjustment where a selection is and an explicit environment × income term, specified a with the interaction interpreted as a substantive rather than as a robustness together, the relational premise has at each of a hypotheses for whom and under what conditions an environment both persons and models and how effects across and conditional on the persons and environments P×E is therefore less a matter of variables than of changing what a study is to be to role of socio-economic status (SES) illustrates why this frequently reports associations between greener neighborhoods and better mental health (e.g., Van den Berg et al., 2015, for an influential review). Yet such direct Environment → Outcome associations may the of individuals are to in green, income as a standard methodological response to this is to income as a statistical covariate, its effect to the contribution of the this is to a better it can be treating income as a source of noise to be subtracted from the we the model and an that does not this approach on a linear that the underlying it to capture the experience of the environment, where the same green space may as a basic survival resource for a low-income resident and a lifestyle among for someone with greater financial toward a transactional perspective, we argue that income should not be treated as a variable to be but as a among of the human-environment Income should be conceptualized as a variable that as a in our It the of environmental affordances available to an individual and the psychological of those from a approach to one that explicitly models the interaction between economic and physical settings, environmental psychology can beyond associations and the relational that how people or within their environments, the complexity of these and which be to reliable outcome modeling these can therefore research beyond psychological outcomes to environmental exposure alone and toward understanding how environmental conditions and social shape are understood as domain-general methodological rather than as a substantive in with existing domain-specific already particular person and environment variables and how they the implications concern how such relations are sampled, and and can therefore be applied within, rather than instead of, these psychology has a of methodological that has its scientific The next is to that to the theoretical of the The distinction between E→PO and P×E should not be as a between does E→PO describe the field as a it a tendency most in research on restorative environments, green space, and environmental Environmental characteristics do psychological experience, and estimating those effects with should be is the reduction of environments to and psychological experience to measurable outcomes of The productive in rigorous measurement with attention to relational The question should not be only what environments do to people, but also what emerges between people and environments and how people and the places they a of environmental change, this perspective becomes increasingly Climate and environmental transformation affect not only physical settings but also the identities, and relationships that people develop with places. these processes requires a discipline capable of psychological phenomena while the complexity that them environmental psychology does not choosing between scientific and conceptual It requires that the most rigorous account of human-environment relationships is one that does not of their inherently relational